Life of Brian Ray

Tonight's meeting (all in Chicago are welcome to attend, new to Python or old) main headline will be "GPU programming with CUDA (http://en.wikipedia.org/wiki/CUDA) and some degree of Python -- Cosmin Stejerean".

This sounds like a great topic. And the venue will be Depaul's CTI on Wabash in the loop.

I went the CAWUG / Cocoa Heads after party on Tuesday. The turn out was great. And it looks like the group does not have an official website and leans toward a Google Group http://groups.google.com/group/cawug organization method. I guess this says something about user group organization not necessarily being tied to the groups lead (Bob Frank) level of effort in putting the event together. There seems to be a more constant venue and speakers panel with CAWUG--a different approach than what is taken by ChiPy.

So, I will be interested in our turn out Today. I think it will be good. I am going to try and get the so-called ChiPy organizers to vote on use of MeetUp and other topics to help the success of the group.

We (the organizers of ChiPy) get a lot of comments on the way we organize. For example, meetings are Just in Time compiled. One issue that makes it less predictable is that we like to move the venue around, a lot. I like moving the venue around. For one thing, it give alternate schools, business, and organizations a chance to host a meeting and develop a new relationship with ChiPy. Likewise it spreads our exposure a bit. I do believe meetings should be in the loop, because its centrally located and accessible form anywhere using mass transit. I also like a mix of big venues and small (sometimes scary) ones. It helps will cohesion of the group members.

My full announcement for today's meeting: http://mail.python.org/pipermail/chicago/2008-July/004504.html

I wrote this script to accomplish the following:

• use an external SMTP email server
  
• Use Python
  
• Not use the svn.fs packages form subversion
  
• Allow configuration on a reg-ex pattern
  
• Send the logs and diffs in the email
  
• allow simply dropping the script into svn/hooks/post-commit
  

First, I tried the contrib scripts from subversion. They were a little too complicated and had dependencies I did not want to install on a shared server.

I wrote this script to accomplish the goals.

Here are some links to the script:

• Plain source
  
• Colorized source
  

Or here is the source inline: Continue reading "Send Custom Email from Subversion Commits Tutorial"

(I also posted this question to the ChiPy List)

I posted the video and will be reposting a better quality version soon
if some of you can help me work out the details (Disclaimer: I am
totally new to recording and publishing video). If you can't wait here
is the first try:



Link to Video

Admittedly, the quality is poor on my first try. The aspect ratio I
shot in was different than how I posted to Google Video--so its skewed
to letterbox. The video looks and sounds much better on my computer
before I post. I shot the video in HD widescreen and went to mp4 the
way the google video suggests in its FAQ. I am not sure I like the
quality of any of the videos on a hosted web viewer solution versus
the pod-casts I view through iTunes. But, I am also concerned if
other pod-cast readers are readily available (Linux? Windows?) so that
all of you viewing can use any feed I produce.

If everyone can use this format, I would like to explore the option of
using the mv4 format in the form of a podcast. IE, it would be nice to
have a RSS feed just for meetings. I did see a Python package called
PyRSS2Gen http://www.dalkescientific.com/Python/PyRSS2Gen.html. Does
anybody know the correlation between the content of the RSS entry for
video and the example used on this website? For instance, is the
content of each entry just html pointing to the video data as a
hyperlink? What do I need to consider if I want to optimize for
devices like the iPhone and iPod?

Also, the video is kind of large in the mv4 or mp4 formats. Its like
around 400-600Meg for 1 hour with the compressions I have available on
my platform. Also, the video when digitized from the HD format takes
60 Gigs. So, I do not really want to save the original data on disk.
Also, I do not really want to host the data in the mv4 or mp4 formats
on the ChiPy Server. Its just a lot of disk space and bandwidth. What
other options do I have. Can I use a file hosting service in
conjunction with a link from a RSS feed I do host on ChiPy Server?

Someone else suggested a torrent. I am not sure how this works either.

Like I mentioned before, I am a novice and do want to continue
recording and sharing this information online. Any suggestions or
advice? I am looking forward to learning but do not know where to start.

My tea sample of "Ancient Pu-Erh Mini Tuo Cha" has the Fair Trade logo on the package. Excellent Tea, by the way; however, what is this Fair Trade all about. I only understood this on very surface level so I looked it up.

I found one of the best wikipedia articles on the topic http://en.wikipedia.org/wiki/Fair_trade. The article highlights the motivations, politics, and history.

The fair trade applies mostly to: "handicrafts, coffee, cocoa, sugar, tea, bananas, honey, cotton, wine, fresh fruit, and so on." Commodities.

As I sip my tea, I watch Maura, my two year old play with toys also from China. Are these toys becoming a commodity much like the sugar and teas, before?

I do not know what organizations make sure toys and clothing are fairly traded. I am sure there are some. But, now is it time to treat these like tea? The data suggests better products and a better world situation that Fair Trade certification helps provide. Its a two way street. Implied by the value in Fair Trade will consumers pay more for a product allowing better understanding on how they were produced. The goal of Fair Trade is socially motivated. With this, we get better products and better understanding of the impact of mass producing an export product. I doubt we will see a Fair Trade logo on products from Walmart or Gap anytime soon. Nonetheless, would it increase the value of these products?

After taking some time off from Django, I have returned and am happy with what I found--in particular "newforms."

The thing I like about Django is it seems to continue to move forward despite some drastic swings in loyal users and a API changes. When a open source project starts to get some attention, I have often found that the projects goal get lost in the clutter by persons involved with mixed motives. The project tear apart in a way not repairable. The project becomes a sort of a Humpty Dumpty. I can tell that by newforms the project continues on track. Yes, its anouther big change for those who can't wait for 1.0, but they seemed to pull it off.

"newforms" is like the removal of a thorn from Django side. The old way of handling forms was one of the things I did not like about Django. It seemed too clumsy. Nevertheless, the newforms method of modeling the intent of a form and then providing a form in table or list item format is fabulous. newforms method of self generation from models is brilliant. The error handling and the round tripping of data is great. "newforms" provides the instant gratification I needed for the one-off project I had to produce under time constraints.

It is really just a problem with the way the path is specified.

Often from my friends in Window's land, I get the paths like:

\\SERVER\share\path\to\file.txt

When converted to a file url this will look something like:

file://SERVER/share/path/to/file.txt

Unfortunately, when I mount the volume on OSX using the normal method. The share showes up under /Volumes with a naming scheme not compatible with the file url above.

My work around involves making a directory and mounting:

1. mount the server the normal way (Pretzel+K in Finder)
2. Check Remember Password option when authenticating
3. (Important) unmount the drive in the finder
4. make a directory with the sever and share name "mkdir -p /SERVER/share"
5. mount the server with the command "mount_smbfs //`whoami`@SERVER/share/ /SERVER/share". NOTE: if your short unix username does not match your windows server login, make it match. If it requires a domain, things like, "NNN\;bray" is perfectly acceptable.

Once this is done, you should test it out and click on the file url. This should bring up the program in the same application if you had double clicked on the file as assigned by Launch Services.

To take it a bit further, you can mount the drive from startup.

1. Create directory "mkdir -p /Library/StartupItems/AutoMount"
2. Make Files AutoMount and StartupParameters.plist in that director

In the AutoMount file put:

#!/bin/sh

. /etc/rc.common

# The start subroutine

StartService() {

   mount_smbfs //NDNW:bray@SERVER/share/

}

# The stop subroutine

StopService() {

  unmount SERVER/share

}

# The restart subroutine

RestartService() {

 StopService

 StartService

}

RunService "$1"

In the StartupParameters.plist goes.

{

  Description     = "Mounts";

  Provides        = ("Mounts");

  OrderPreference = "Late";

  Messages =

  {

    start = "Starting Software Update service";

    stop  = "Stopping Software Update service";

  };

}

You should restart and it should mount automatically. If you do not see it, look in your console for notes. I tested the method on 10.4, but should work on 10.5, as well.

If you want to be technical or you just like to read pointless crap you may find more details from apple's site here http://developer.apple.com/documentation/MacOSX/Conceptual/BPSystemStartup/Articles/StartupItems.html

Our MoinMoin wiki, found at http://chipy.org has been getting some spam even with MoinMoin's Anti-Spam features turned on. At least, that is, until I did something a week or so ago to plug this up. Its not that I haven't tried to do something in the past: blocked IPs, blocked urls, LocalBadContent, and allowed users to delete pages... I may be inviting more spammers by blogging about this, but I would like to share my resolution.

I thought of using an image Captcha. It worked for my own contact form. But there is something about accessibility and future venerabilities to OCR hacks, that make me wonder if this is the best solution.

Another approach, was to work with one of the SpamBayes authors to make a MoinMoin solution. I would be still open to this, but it may require more participation on my part than I have time for right now.

The solution I came up with was a Question if the form (Question, Correct Answer1, Correct Answer2):

"What is one plus one:","2","two"
"What is the color of the sky:","blue","blue"
"What is the color of grass:","green","green"
"How many inches are there in a foot:","twelve","12"
"The opposite of big:","small","tiny"
"The opposite of small:","big","large"
"What city is known as The Windy City:","chicago","chitown"
"Which way is the sky:","up","up"
"Which direction is the North Poll, (north, west, south, or east):","north","N"
"ChiPy mascot is a chipmunk, 'Y' or 'N', :","y","yes"
"The world is Flat, 'Y' or 'N':","n","no"

Its not the first time I have seen this. Ian or someone said that they did this with limited success. There has been some talk on the MoinMoin lists of users complaining about the same thing. I do not want to have to keep a list (an ACL or wikipage) of those who can post. I just do not like this solution. Esp. when its getting close to meeting time and a new poster needs access.

So it works like this. Since I can generate the JavaScript Server side, I give one of the questions from above, just one. On is pulled in from randomly from Python. Once a user posts, an "onClick()" event is sent. This brings up a choice JavaScript dialog with the question. The Answer and the Question are sent to a hidden form field then sent trough to the POST. Additionally, to avoid robots, the answer is wrapped with a string that changes daily (unfortunate to those who start writing before midnight and finish after midnight). The string is stripped out and answer is checked on client side.

Feel free to test: http://www.chipy.org/EditTest.

(Here is the patch to MoinMoin, I do not know what version)

A Python/C++ project I am working on requires a lightweight human readable/writable markup. Unless you stop me, I am about to pick YAML...

Continue reading "Is picking YAML a bad thing?"

Go Cubs!

So to be able to rant a bit more about purely Print-Technology items and not bore people who read this blog, I have started a new group blog entitled "Life of Print." Its located here: Life of Print. Continue reading "I started the "Life of Print," Blog"

I quit my job working as software engineer for Printable Technologies. I love those guys and the clients--I am going to miss you all. I have started something really big and exciting I can not talk about yet.. stay tuned I will post more about it soon. No, I am not leaving the print industry.

Since I know many who read this blog are from the Mac Dev world, I thought that this would be a good place to post a couple links to the jobs we have posted on Craig's List:

http://chicago.craigslist.org/chc/sof/421418294.html

http://chicago.craigslist.org/chc/sof/421412026.html

BTW, I did not write these descriptions, so do not complain to me about the misuse of terms. Nevertheless, I can vouch this is a good job with a good company.

Trolling for Python news, I ran across this article Make Your Pilgrimage to Python Bootcamp at the Big Nerd Ranch, November 5-9, 2007. It states that the class teaches, "using Python to process data, creating links between existing applications, and creating complex mission critical applications involving concurrency, network programming, and distributed systems." Better yet, the class will be thought by David Beazley, one of the great Python Guru's also from Chicago. Seems to be a really good choice for Big Nerd Ranch Class. Here is the complete press release:

Continue reading "Python Bootcamp taught David Beazley"

The FSF GPLv3 is not so much more radical over its predecessors. Technically, its not backward compatible with V2. But, still, its clear to those who know what was intended; what's intended and implied is not just the copy leftness, but the underlaying philosophy of being able to replace a piece of GPL'd technology already should have covered "Installation information." The problem still exists for embedded technology, as it did before. Continue reading "GPLv3 installation information on embedded technology"